Research: Overall Server Shipments to Grow 13% Annually, General-Purpose Server Lead Times Extended
TrendForce predicts that in 2026, demand for both general-purpose and AI servers will increase. However, suppliers are prioritizing production capacity for higher-profit AI servers, significantly extending lead times for general-purpose server components. This will limit overall server shipment growth to around 13% annually, failing to fully reflect potential market purchasing power. Lead times for power ICs and BMC ICs are particularly extended, and Samsung's plan to close its 8-inch wafer fab in Korea will further squeeze PMIC supply for general-purpose servers.
📋 Article Processing Timeline
- 📰 Published: April 15, 2026 at 17:55
- 🔍 Collected: April 15, 2026 at 18:01 (6 min after Published)
- 🤖 AI Analyzed: April 15, 2026 at 19:38 (1h 36m after Collected)
Taipei, April 15 (CNA) Research firm TrendForce stated that in 2026, AI will simultaneously drive demand for general-purpose servers and AI servers. However, due to suppliers prioritizing production capacity for higher-profit AI servers, lead times for general-purpose server components have significantly lengthened. The overall server shipment growth, originally expected to be nearly 20% annually, will thus remain around 13%, unable to fully reflect the market's potential purchasing power.
TrendForce pointed out in its latest server industry research that demand for general-purpose servers remains steady but has already faced tight supply of printed circuit boards (PCBs) and central processing units (CPUs). Currently, lead times for these two core components have extended to nearly a year. In addition, lead times for power ICs and BMC (Baseboard Management Controller) ICs have also significantly lengthened recently.
Regarding power management ICs (PMICs), because AI servers require much higher power density than general-purpose servers and are prioritized by suppliers, 8-inch wafer BCD processes are heavily skewed towards AI power management chips. To make matters worse, Samsung's plan to close its S7 8-inch wafer fab in Korea will further squeeze PMIC production capacity for general-purpose servers, extending lead times from 21-26 weeks to 35-40 weeks.
For AI servers, strong demand from cloud service providers (CSPs) is expected to support an annual shipment growth of approximately 28% in 2026. It is anticipated that application-specific integrated circuit (ASIC) AI servers will see greater shipment growth than graphics processing unit (GPU) AI servers. However, considering the time-consuming tuning of self-developed chips by companies like Meta and AWS, there is a risk of shipment delays. TrendForce has slightly adjusted the proportion of ASIC AI servers in overall AI servers for 2026 from nearly 28% to around 27%, with GPU models maintaining the majority. (Editor: Yang Lan-hsuan) 1150415
Stand with the facts, your sponsorship is the power to protect press freedom.
Download CNA's "First-hand News" APP to stay updated with the latest news.
The text, images, and videos on this website may not be reproduced, publicly broadcast, publicly transmitted, or utilized without authorization.
TrendForce pointed out in its latest server industry research that demand for general-purpose servers remains steady but has already faced tight supply of printed circuit boards (PCBs) and central processing units (CPUs). Currently, lead times for these two core components have extended to nearly a year. In addition, lead times for power ICs and BMC (Baseboard Management Controller) ICs have also significantly lengthened recently.
Regarding power management ICs (PMICs), because AI servers require much higher power density than general-purpose servers and are prioritized by suppliers, 8-inch wafer BCD processes are heavily skewed towards AI power management chips. To make matters worse, Samsung's plan to close its S7 8-inch wafer fab in Korea will further squeeze PMIC production capacity for general-purpose servers, extending lead times from 21-26 weeks to 35-40 weeks.
For AI servers, strong demand from cloud service providers (CSPs) is expected to support an annual shipment growth of approximately 28% in 2026. It is anticipated that application-specific integrated circuit (ASIC) AI servers will see greater shipment growth than graphics processing unit (GPU) AI servers. However, considering the time-consuming tuning of self-developed chips by companies like Meta and AWS, there is a risk of shipment delays. TrendForce has slightly adjusted the proportion of ASIC AI servers in overall AI servers for 2026 from nearly 28% to around 27%, with GPU models maintaining the majority. (Editor: Yang Lan-hsuan) 1150415
Stand with the facts, your sponsorship is the power to protect press freedom.
Download CNA's "First-hand News" APP to stay updated with the latest news.
The text, images, and videos on this website may not be reproduced, publicly broadcast, publicly transmitted, or utilized without authorization.
FAQ
What are the main trends in the server market for 2026?
In 2026, demand for both AI and general-purpose servers is expected to rise. However, due to production capacity prioritization for AI servers, lead times for general-purpose server components will extend, limiting overall shipment growth to about 13%.
What are the primary reasons for extended lead times for general-purpose server components?
Suppliers are prioritizing higher-profit AI servers, and there's a tight supply of key components like PCBs, CPUs, power ICs, and BMC ICs. Samsung's plan to close its 8-inch wafer fab also contributes to the issue.